Latest news with #WeiSun
CNBC
29-04-2025
- Business
- CNBC
Alibaba launches new Qwen LLMs in China's latest open-source AI breakthrough
Alibaba released the next generation of its open-sourced large language models, Qwen3, on Tuesday — and experts are calling it yet another breakthrough in China's booming open-source artificial intelligence space. In a blog post, the Chinese tech giant said Qwen3 promises improvements in reasoning, instruction following, tool usage and multilingual tasks, rivaling other top-tier models such as DeepSeek's R1 in several industry benchmarks. The LLM series includes eight variations that span a range of architectures and sizes, offering developers flexibility when using Qwen to build AI applications for edge devices like mobile phones. Qwen3 is also Alibaba's debut into so-called "hybrid reasoning models," which it says combines traditional LLM capabilities with "advanced, dynamic reasoning." According to Alibaba, such models can seamlessly transition between a "thinking mode" for complex tasks such as coding and a "non-thinking mode" for faster, general-purpose responses. "Notably, the Qwen3-235B-A22B MoE model significantly lowers deployment costs compared to other state-of-the-art models, reinforcing Alibaba's commitment to accessible, high-performance AI," Alibaba said. The new models are already freely available for individual users on platforms like Hugging Face and GitHub, as well as Alibaba Cloud's web interface. Qwen3 is also being used to power Alibaba's AI assistant, Quark. AI analysts told CNBC that the Qwen3 represents a serious challenge to Alibaba's counterparts in China, as well as industry leaders in the U.S. In a statement to CNBC, Wei Sun, principal analyst of artificial intelligence at Counterpoint Research, said the Qwen3 series is a "significant breakthrough—not just for its best-in-class performance" but also for several features that point to the "application potential of the models." Those features include Qwen3's hybrid thinking mode, its multilingual support covering 119 languages and dialects and its open-source availability, Sun added. Open-source software generally refers to software in which the source code is made freely available on the web for possible modification and redistribution. At the start of this year, DeepSeek's open-sourced R1 model rocked the AI world and quickly became a catalyst for China's AI space and open-source model adoption. "Alibaba's release of the Qwen 3 series further underscores the strong capabilities of Chinese labs to develop highly competitive, innovative, and open-source models, despite mounting pressure from tightened U.S. export controls," said Ray Wang, a Washington-based analyst focusing on U.S.-China economic and technology competition. According to Alibaba, Qwen has already become one of the world's most widely adopted open-source AI model series, attracting over 300 million downloads worldwide and more than 100,000 derivative models on Hugging Face. Wang said that this adoption could continue with Qwen3, adding that its performance claims may make it the best open-source model globally — though still behind the world's most cutting-edge models like OpenAI's o3 and o4-mini. Chinese competitors like Baidu have also rushed to release new AI models after the emergence of DeepSeek, including making plans to shift toward a more open-source business model. Meanwhile, Reuters reported in February that DeepSeek is accelerating the launch of its successor to its R1, citing anonymous sources. "In the broader context of the U.S.-China AI race, the gap between American and Chinese labs has narrowed—likely to a few months, and some might argue, even to just weeks," Wang said. "With the latest release of Qwen 3 and the upcoming launch of DeepSeek's R2, this gap is unlikely to widen—and may even continue to shrink."
Yahoo
28-04-2025
- Business
- Yahoo
DeepSeek is hiring for an 'urgent' role in product management and design
DeepSeek is hiring for a job in product management and design. It's a major shift from the startup's focus on AI model research. The rush to hire product talent mirrors a broader trend in the US. DeepSeek, the Chinese startup that rattled the AI industry earlier this year, is hiring for a product role that illustrates the company's shift from research to commercialization. In a job notice posted Tuesday on its official WeChat account, DeepSeek said it is looking to fill a "product and design" position on its teams in Beijing and Hangzhou. It is unclear from the notice if the job refers to a single role or multiple positions. The Hangzhou-based firm labeled the job notice "urgent." The company wrote that it wants people to help create the "next generation of intelligent product experience" centered on large language models. Candidates are expected to have product management experience and be proficient in product and visual design, the notice said. DeepSeek did not respond to Business Insider's request for comment. DeepSeek is also hiring a chief financial officer and chief operating officer — jobs not labeled urgent. The company is expanding its research and engineering teams, according to other listings on its WeChat account. The move marks a major shift for the company, which has been focused on fundamental AI model research. Last month, DeepSeek released an upgraded version of its open-source V3 large language model, boosting its reasoning and coding capabilities. Founded in 2023 by Chinese entrepreneur Liang Wenfeng, DeepSeek made headlines and disrupted markets in January after unveiling its low-cost reasoning model, R1. The startup claims R1 can rival top competitors like OpenAI's GPT-4 — but at a fraction of the cost. An analyst told Business Insider earlier this month that DeepSeek's latest models — especially the reasoning-focused R1 and R2 set to launch later this month or in May — mark a "significant inflection point" in China's AI ambitions. "These models not only match the best-in-class performance globally, but are also open-sourced under the most permissive MIT License," said Wei Sun, the principal analyst for AI at Counterpoint Research. "That changes the game," she added. Unlike flagship models in the US, which are typically closed-sourced and monetized through APIs or enterprise licensing, DeepSeek's models like R1 and V3 are free for anyone to download, modify, and integrate. DeepSeek has been quiet about the progress of its next-generation R2 model. Amid high costs and chip shortages, Chinese firms have been prioritizing AI integration and consolidation to stay competitive, an analyst told BI earlier this month. Tencent has deployed its Hunyuan model and DeepSeek R1 across its massive ecosystem, including WeChat, said Ray Wang, a Washington-based analyst who specializes in AI and US-China tech statecraft. WeChat, China's biggest social media app, is used by nearly 1.4 billion people. Baidu has also integrated DeepSeek R1 into its search engine, he said. While details about DeepSeek's hiring process are scant, Liang, the founder, has made it clear that he values creativity over experience when it comes to hiring. In a 2023 interview with Chinese tech publication 36KR, he said that "experience is not that important," even in a similar role. "Basic skills, creativity, and passion are much more important," he added. "Our core technical positions are mainly filled by fresh graduates or those who have graduated one or two years ago," he said. The rush to hire product talent mirrors a broader trend in the AI world. In the US, product managers are seen as increasingly critical for some companies in the AI era, helping bridge the gap between rapidly advancing AI technology and real-world user needs. "The future really does belong to product managers," Frank Fusco, a product manager turned CEO of a software company called Silicon Society, told BI in November. As AI becomes more capable of handling coding and other engineering tasks, Fusco said it's an opportunity for product managers to take on an even greater role. OpenAI is hiring seven product manager roles in its New York and San Francisco offices, and Anthropic is hiring 11 product-related roles, according to the companies' websites. However, some tech companies are revisiting their reliance on product managers. Microsoft wants to increase the number of engineers relative to product or program managers, BI's Ashley Stewart reported last month. Other companies like Airbnb and Snap have been rethinking the need for product managers. Read the original article on Business Insider Sign in to access your portfolio
Business Insider
28-04-2025
- Business
- Business Insider
DeepSeek is hiring for an 'urgent' role in product management and design
DeepSeek, the Chinese startup that rattled the AI industry earlier this year, is hiring for a product role that illustrates the company's shift from research to commercialization. In a job notice posted Tuesday on its official WeChat account, DeepSeek said it is looking to fill a "product and design" position on its teams in Beijing and Hangzhou. It is unclear from the notice if the job refers to a single role or multiple positions. The Hangzhou-based firm labeled the job notice "urgent." The company wrote that it wants people to help create the "next generation of intelligent product experience" centered on large language models. Candidates are expected to have product management experience and be proficient in product and visual design, the notice said. DeepSeek did not respond to Business Insider's request for comment. DeepSeek is also hiring a chief financial officer and chief operating officer — jobs not labeled urgent. The company is expanding its research and engineering teams, according to other listings on its WeChat account. The move marks a major shift for the company, which has been focused on fundamental AI model research. Last month, DeepSeek released an upgraded version of its open-source V3 large language model, boosting its reasoning and coding capabilities. Founded in 2023 by Chinese entrepreneur Liang Wenfeng, DeepSeek made headlines and disrupted markets in January after unveiling its low-cost reasoning model, R1. The startup claims R1 can rival top competitors like OpenAI's GPT-4 — but at a fraction of the cost. An analyst told Business Insider earlier this month that DeepSeek's latest models — especially the reasoning-focused R1 and R2 set to launch later this month or in May — mark a "significant inflection point" in China's AI ambitions. "These models not only match the best-in-class performance globally, but are also open-sourced under the most permissive MIT License," said Wei Sun, the principal analyst for AI at Counterpoint Research. "That changes the game," she added. Unlike flagship models in the US, which are typically closed-sourced and monetized through APIs or enterprise licensing, DeepSeek's models like R1 and V3 are free for anyone to download, modify, and integrate. DeepSeek has been quiet about the progress of its next-generation R2 model. Amid high costs and chip shortages, Chinese firms have been prioritizing AI integration and consolidation to stay competitive, an analyst told BI earlier this month. Tencent has deployed its Hunyuan model and DeepSeek R1 across its massive ecosystem, including WeChat, said Ray Wang, a Washington-based analyst who specializes in AI and US-China tech statecraft. WeChat, China's biggest social media app, is used by nearly 1.4 billion people. Baidu has also integrated DeepSeek R1 into its search engine, he said. While details about DeepSeek's hiring process are scant, Liang, the founder, has made it clear that he values creativity over experience when it comes to hiring. In a 2023 interview with Chinese tech publication 36KR, he said that "experience is not that important," even in a similar role. "Basic skills, creativity, and passion are much more important," he added. "Our core technical positions are mainly filled by fresh graduates or those who have graduated one or two years ago," he said. Rise of product managers The rush to hire product talent mirrors a broader trend in the AI world. In the US, product managers are seen as increasingly critical for some companies in the AI era, helping bridge the gap between rapidly advancing AI technology and real-world user needs. "The future really does belong to product managers," Frank Fusco, a product manager turned CEO of a software company called Silicon Society, told BI in November. As AI becomes more capable of handling coding and other engineering tasks, Fusco said it's an opportunity for product managers to take on an even greater role. OpenAI is hiring seven product manager roles in its New York and San Francisco offices, and Anthropic is hiring 11 product-related roles, according to the companies' websites. However, some tech companies are revisiting their reliance on product managers. Microsoft wants to increase the number of engineers relative to product or program managers, BI's Ashley Stewart reported last month. Other companies like Airbnb and Snap have been rethinking the need for product managers.
New York Times
11-04-2025
- Business
- New York Times
How TikTok's Parent, ByteDance, Became an A.I. Powerhouse
The Chinese internet giant ByteDance has made some of the world's most popular apps: TikTok and, in China, Douyin and Toutiao. In the United States, TikTok claims 170 million users. But in China, about 700 million use the domestic version, Douyin, and 300 million scroll the headlines on Toutiao, a news app. Every video that ByteDance's users watch or post gives the company another data point about how people use the internet. For years, ByteDance has applied that wealth of information to make its apps more appealing, improving its ability to recommend content to keep users hooked. ByteDance is also using the data as the linchpin of a growing business in artificial intelligence. The company has invested billions of dollars in the infrastructure needed to power A.I. systems, building vast data centers in China and Southeast Asia and buying up advanced semiconductors. ByteDance is also on an A.I. hiring spree. ByteDance is best known outside China for TikTok, an app so popular that at least 20 governments have adopted partial bans over concerns about its influence on national security and public opinion. Concern over how ByteDance uses data has driven lawmakers in Washington to try to force a sale of TikTok's U.S. operations. On Friday, President Trump extended a looming deadline by 75 days into mid-June. But in China all that data has helped ByteDance expand its business far beyond social media and gain an edge in the global race to build advanced A.I. technology. 'ByteDance has all this data, all the time, from millions of users,' said Wei Sun, a principal analyst in artificial intelligence at Counterpoint Research in Beijing. Officials in Beijing have pushed China's tech companies to pivot from entertainment apps to what the government sees as an existential goal: self-reliance in cutting-edge technologies that also have military applications, like semiconductors, supercomputers and artificial intelligence. ByteDance has embraced that mission. Last year, the company spent roughly $11 billion on infrastructure like data centers, networking equipment and computer chips, according to a report by Zheshang Securities, a Chinese financial firm. The Biden administration set up rules to try to keep Chinese companies from getting access to those kinds of chips, particularly ones made by Nvidia, the Silicon Valley giant. But ByteDance has found ways to get the computing power it needs to train its systems — in part by using data centers outside China and most likely, analysts say, by buying chips made by Chinese chipmakers like Huawei and Cambricon. While these Chinese-made chips cannot do everything the Nvidia chips can do, they work well enough to help companies like ByteDance provide A.I. services to people and businesses in China. Chinese tech companies have been 'encouraged to adopt local options' for buying chips, said Lian Jye Su, an analyst at Omdia, a market research firm. All this spending has helped ByteDance make one of the most popular artificial intelligence apps in China. Its chatbot, Doubao, gained 60 million users within its first three months on the market last year. It was China's most popular chatbot, beating rivals made by Baidu and Alibaba-backed Moonshot, until the start-up DeepSeek released its own this year. ByteDance showed how closely connected its app ecosystem is with its A.I. efforts when it recently started allowing some users to chat with Doubao inside the Douyin app. In 2021, ByteDance started Volcano Engine, a business that lets other companies pay to use the technologies that made TikTok, Douyin and Toutiao so addictive, like tools to analyze information and the algorithms that recommend videos. Some of these services were natural applications of the technology that ByteDance developed for Douyin and TikTok, like filters that can make people appear much older or superimpose sparkly hearts on their faces. ByteDance used its experience making these filters to help companies like Haier and Hisense develop movement-tracking technology for gesture-controlled home appliances like smart televisions. GAC Group, one of China's largest makers of electric vehicles, is using Volcano Engine to translate and manage data for cars sold outside China. And Mercedes-Benz said last year that it would use Volcano Engine in its in-car voice assistant and navigation system in China. ByteDance did not respond to a request for comment. Company job postings show that ByteDance is hiring for hundreds of A.I.-related roles. The company recently directed its engineering team to focus on a milestone that tech companies like OpenAI, Google and DeepSeek are also chasing — making an A.I. system that is as smart as or smarter than humans, often referred to as artificial general intelligence. While many Chinese companies have started A.I. projects, a much smaller number have the resources to invest in the personnel and computing power needed to advance the technology. Some experts expect that a research team somewhere in the world will make this kind of system within the next year or two.
Yahoo
09-04-2025
- Business
- Yahoo
Lumentum Holdings Inc. (LITE) Debuts 400G PAM4 Optics to Boost AI and Cloud Networking Speeds
We recently published a list of . In this article, we are going to take a look at where Lumentum Holdings Inc. (NASDAQ:LITE) stands against other AI news updates on Wall Street's radar. China has sensed a window of opportunity in its race to usurp US dominance in AI innovation. As Washington remains focused on building regulatory walls that restrict access to AI, Beijing has resorted to fostering open-sourcing artificial intelligence models to boost AI innovation and adoption. That was clear when DeepSeek unveiled cost-effective AI models that challenged American tech dominance on large language models and data centers. While the initial focus was on how DeepSeek came up with powerful AI models that rivaled US models, the focus has since changed to how the Chinese startup catalyzed the adoption of open-source AI models. The decision to open source AI source code comes as Chinese companies look to capitalize on growing opportunities in the artificial intelligence software market, growing at a compound annual growth rate of 35.52% and expected to be worth $223.35 billion by 2028, according to Research and Markets. 'DeepSeek's success proves that open-source strategies can lead to faster innovation and broad adoption,' said Wei Sun, principal analyst of artificial intelligence at Counterpoint Research. Understanding that it will require time for China to reach the level of the US in advanced AI computing, Chinese firms have concentrated on creating more efficient and cost-effective AI solutions. They are also striving to establish leading roles in open-source AI, cloud services, and worldwide data networks. These approaches allow China to provide more affordable and unrestricted AI access to nations discontented with US policies, integrating it into developing markets in ways that will be challenging to disrupt. This is not merely a contest in AI; it is a struggle for dominance over the global digital infrastructure both now and in the future. Since DeepSeek sent shockwaves, several Chinese companies have released open-source models that are free for individual users. The move to open-source AI models underscores a broader shift in China that no longer focuses on proprietary licensing. Instead, Chinese companies freely offer the underlying source code for modification and redistribution. While US tech giants have always insisted on open-sourcing their models or source code, that has not always been the case in entirety, for starters, most US companies claim to use open-source code while still restricting their use and modifications. Other US companies generate revenue by charging people to access their models even on claiming they are open source. For this article, we selected AI stocks by going through news articles, stock analysis, and press releases. These stocks are also popular among hedge funds in Q4 2024. Why are we interested in the stocks that hedge funds pile into? The reason is simple: our research has shown that we can outperform the market by imitating the top stock picks of the best hedge funds. Our quarterly newsletter's strategy selects 14 small-cap and large-cap stocks every quarter and has returned 373.4% since May 2014, beating its benchmark by 218 percentage points (). A close-up of a technician calibrating a laser beam. Number of Hedge Fund Holders: 57 Lumentum Holdings Inc. (NASDAQ:LITE) is a leading developer and provider of optical and photonic products for the cloud and networking markets. On March 31, the company affirmed the significant milestone it has achieved in advancing next-generation, high bandwidth, and lower power per bit optics for artificial intelligence (AI)/machine learning (ML) and cloud infrastructure. In collaboration with Marvell Technology, a leader in data infrastructure, Lumentum Holdings Inc. (NASDAQ:LITE) is poised to showcase the industry's first 400G/per lane PAM4 technology operating at 225 Gbaud. The live demonstration will feature a cutting-edge 450G PAM4 distributed feedback (DFB) laser integrated with a Mach-Zehnder (MZI) modulator. The exhibition highlights the preparedness for next-generation 3.2T optical transceivers, which are crucial for the swiftly growing AI/ML tasks and large-scale cloud networking frameworks. 'Lumentum's collaboration with Marvell marks our joint commitment to pushing the boundaries of 400G per lane technology innovation,' said Matt Sysak, Lumentum CTO of Cloud and Networking. 'Together, we are enabling 400G per lane solutions that go beyond what silicon photonics can offer, unlocking higher performance and scalability foundational for next-generation AI/ML and cloud infrastructure.' Overall, LITE ranks 3rd on our list of AI news updates on Wall Street's radar. While we acknowledge the potential of LITE as an investment, our conviction lies in the belief that AI stocks hold greater promise for delivering higher returns and doing so within a shorter time frame. There is an AI stock that went up since the beginning of 2025, while popular AI stocks lost around 25%. If you are looking for an AI stock that is more promising than LITE but that trades at less than 5 times its earnings, check out our report about the . READ NEXT: 20 Best AI Stocks To Buy Now and 30 Best Stocks to Buy Now According to Billionaires. Disclosure: None. This article is originally published at Insider Monkey. Sign in to access your portfolio
